Comprehensive analysis of gene expression is rapidly becoming a possibility not only for selected model systems, but also for clinical tissue samples. The analysis of gene expression in clinical tissue samples has been severely limited, however, by the standard practice of formalin-based fixation in surgical pathology. Formalin impacts the quality and quantity of cellular macromolecules that can be extracted from fixed tissue. Therefore, any quantitative analysis has historically depended on excess tissue, which could be obtained without compromising the pathologic evaluation. As a result, there is almost no material available from small samples, thus effectively precluding analysis of precancerous lesions and early invasive cancer, both of which may hold the key to many basic questions about the early carcinogenic process. Paradoxically, we are experiencing an unprecedented increase in the accrual of these small lesions thanks to increasing successes in screening programs for both breast and prostate cancer. Clearly, making these specimens available to molecular methods presents a tremendous opportunity to advance our knowledge of the initial phases of carcinogenesis. We propose to address this issue with two complementary approaches. First, improve the currently insufficient methods of RNA and DNA extraction from archival formalin fixed material. Nucleic acid losses due to non-specific trapping in irreversibly crosslinked macromolecules will be minimized by optimizing the breakdown of all non-target components as well as using the affinity of nucleic acids for silica beads in high salt solutions. The template function of extracted nucleic acids will be improved by removing labile N-methylol adducts from the purified polynucleotides. Second, develop new fixation procedures improving the extractability of amplifyable DNA and RNA templates, while avoiding the difficulties of having to divide limited tissue samples according to various clinical and research needs. We have just completed a pilot project demonstrating that alternative alcohol based fixation produces excellent histopathological and immunohistochemical results without compromising extraction of functional messenger RNA, as assessed by quantitative real-time RT-PCR. Fixation without crosslinking may fulfill both clinical and scientific needs.